Mathematicians Hate AI. They Can't Quit It
Wired AI · rss · 2026-09-19
Wired explores the math community's conflicted relationship with AI: powerful models pose an existential risk to the field—from hard-to-verify proofs to eroding independent reasoning—yet researchers can't stop relying on them because they're simply too useful for exploring ideas and accelerating research.
More from AGI Musings
- deepfates attends the Eleos conference on AI consciousness and welfare this weekend — PeterBowdenLive · 2026-09-19
- Why doesn't AI have user-protective 'safety labels' like food? A Reddit proposal — Fun-Engineering6273 · 2026-09-19
- AI researcher fires at Anthropic: LLMs are state machines, anthropomorphism is alchemy — gerardsans · 2026-09-19
- There is no degree collapse, argues exec: lifelong learning beats a 4-year degree — deepakns · 2026-09-19
- When 'why are you talking about this?' is your best rebuttal, Cofnas wins more rationalists — LucaAmb · 2026-09-19
- Jan Kulveit Pushes Back on 'Context Is the Right AI Identity' Claim — gleech · 2026-09-19